* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download The Concepts of Truth and Meaning in the Buddhist Scriptures
Buddhist cosmology wikipedia , lookup
Buddhist cosmology of the Theravada school wikipedia , lookup
Faith in Buddhism wikipedia , lookup
Four Noble Truths wikipedia , lookup
Nirvana (Buddhism) wikipedia , lookup
Relics associated with Buddha wikipedia , lookup
Decline of Buddhism in the Indian subcontinent wikipedia , lookup
Persecution of Buddhists wikipedia , lookup
Buddhist art wikipedia , lookup
Early Buddhist schools wikipedia , lookup
Wat Phra Kaew wikipedia , lookup
Buddhism and psychology wikipedia , lookup
History of Buddhism wikipedia , lookup
Yiqiejing yinyi (Xuanying) wikipedia , lookup
Triratna Buddhist Community wikipedia , lookup
Silk Road transmission of Buddhism wikipedia , lookup
Pratītyasamutpāda wikipedia , lookup
Buddhism and sexual orientation wikipedia , lookup
Dhyāna in Buddhism wikipedia , lookup
Gautama Buddha wikipedia , lookup
Buddhism in Myanmar wikipedia , lookup
Enlightenment in Buddhism wikipedia , lookup
Greco-Buddhism wikipedia , lookup
Buddhist ethics wikipedia , lookup
Women in Buddhism wikipedia , lookup
Buddha-nature wikipedia , lookup
Buddhism and Western philosophy wikipedia , lookup
Sanghyang Adi Buddha wikipedia , lookup
Pre-sectarian Buddhism wikipedia , lookup
Mahayana sutras wikipedia , lookup
Buddhist texts wikipedia , lookup
THE JOURNAL OF THE INTERNATIONAL ASSOCIATION OF BUDDHIST STUDIES EDITOR-IN-CHIEF A. K. Narain University of Wisconsin, Madison, USA EDITORS Heinz Bechert Universitdt Gottingen, FRG Leon Hurvitz UBC, Vancouver, Canada Lewis Lancaster University of California, Berkeley, USA Alexander W. MacDonald Universite de Paris X, Nanterre, France B. f. Stavisky WNIIR, Moscow, USSR Alex Wayman Columbia University, New York, USA ASSOCIATE EDITOR Stephan Beyer University of Wisconsin, Madison, USA Volume 4 1981 Number I CONTENTS I. ARTICLES The Concepts of Truth and Meaning in the Buddhist Scriptures, by Jose I. Cabezon Changing the Female Body: Wise Women and the Bodhisattva Career in Some Maharatnakvtasutras, by Nancy Schuster liodhi and Amhattaphala. From early Buddhism to early Mahayana, h Karel Werner 7 24 70 II. SHORT PAPERS A Study on the Madhyamika Method of Refutation and Its Influence on Buddhist Logic, by Shohei Ichimura 87 An Exceptional Group of Painted Buddha Figures at Ajanta, fry Anand Krishna 96 Rune E. A. Johansson's Analysis of Citta: A Criticism, by Arvind Sharma 101 III. BOOK REVIEWS AND NOTICES Cross Currents in Early Buddhism, fry S. N. Dube 108 Buddha's Lions—The Lives of the Eighty-Four Siddhas, by James B. Robinson 111 Tangles and Webs, by Padmasiri de Silva Buddhist and Freudian Psychology, by Padmasiri de Silva Buddhist-Christian Empathy, by Joseph J. Spae T h e Religions of Tibet, by Guiseppe Tucci 113 114 115 119 IV. NOTES AND NEWS A Report on the 3rd Conference of the IABS 123 Buddhism and Music 127 Presidential Address at the 2nd IABS Conference at Nalanda hyP.Prndhan 128 contributors 143 T h e Concepts of Truth and Meaning in the Buddhist Scriptures by Jose I. Cabezon In 1976 during a visit to a Buddhist monastery on the East coast of the United States I made the acquaintance of a monk of the Theravada tradition. During a series of often heated discussions which ensued my colleague raised this most fascinating and indeed insightful objection. He said to me: "You see the problem is really quite simple, the Hinayana asserts that all of Buddha's words are true while the Mahayana claims that all that is true is the word of the Buddha (huddhavacanam)." T h e claim is a bit facile and by far an overstatement of the situation. T h e more I pondered the problem however, the less offensive I realized a Mayayanist would find it, and in the end I felt that a Mahayanist should feel quite at ease in conceiving of "the set of all truths" as being at least "the intent of the Buddha" (if not his actual words). This was a position which I thought should be perfectly acceptable. In the years that have passed since this occasion, I have steadily pursued my interests in this question. In particular, I have attempted to determine what the Tibetan sources have to say in this regard, and whether it is a consistent account. This brief paper is then the result of some of these investigations.' Scripture and Pramana Buddhism has often been regarded as a non-dogmatic religion, and rightfully so. Despite the claims of some scholars 2 , the critical spirit, so eloquently captured in the parable of the goldsmith, is simply too important and all-pervasive a part of both Hinayana and Mahayana Buddhism to be challenged. 7 Traditionally Buddhism has posited two forms of pramdna: pratyaksa, or direct perception, and anumdna, or inference. In general Buddhists have, with the Vaisesikas and against the Advaitins and Mimamsakas, rejected the validity of sabda (Tib. lung or sgra), scriptural testimony, as a valid source of knowledge. But this must be qualified, for scriptural evidence is, at least according to some Buddhist sources, acceptable with a proviso. Dharmakirti makes the following statement in Pramdnavdrttikain I, 216: Reliable words are non-mistaken. They are a form of inductive inference. 3 dptavdddvisamvdda sdmdnydd anwndnatd4 yid ches tshig ni mi slu ba'i spyi las rjes su dpag pa nyid* Two questions come to mind: (1) what characterizes reliable words and (2) why are they a form of inductive inference? In succeeding verses Dharmakirti explains that for a scripture to be considered reliable (and hence non-mistaken) it must at least not contradict direct perception (pratyaksa) and inference (anumdna). Now rGyal tshab Dar ma rin-chen comments extensively on this point in the Thar lam gsal byed6, his monumental commentary on the Pramdnavdrttikain. He concludes that only as regards very subtle points of doctrine (Tib. shin tu Ikog gyur) can scripture be relied upon as an authority, and that this can be justified inductively. He says that the fact that the less subtle points of doctrine can be (either logically or perceptually) verified to be correct leads one to infer inductively that the very subtle points are also accurate. 7 Moreover only those scriptures which are "purified by means of the three forms of examination" (spyadpa gsum gyis dag pa), i.e., which meet the following three conditions, can be considered authoritative. The three criteria are: 1. that the scripture not contradict the testimony of direct perception (tib. mngon sum la mi grind pa) 2. that it not contradict inferential reasoning (tib. rjes dpag la mi griod pa) and 3. that it not contradict inference based on reliable words (tib. yid ches rjes dpag la mi gnod pa) Now the first two constraints assure us that by taking scripture as valid testimony we are in facr not departing from the conviction that 8 pratyaksa and anumdna are the guiding principles as regards validity. T h e third category is puzzling, for it seemingly involves us in a circular definition by positing reliability as a criterion for a reliable scripture. But this is in fact not. the case. In discussing the third point rGyal tshab gives the following interpretation. He asserts that as a third constraint imposed on valid scriptural testimony, the work in question must be consistent. It cannot contradict other points of scripture, either explicit or implicit (tib. dngos shugsfi? Given the abundance of contradictory statements in the corpus of Buddhist exegesis, this is indeed a rigorous constraint. In short, iabda had to meet very rigorous conditions in order to be considered valid, conditions which most interesting scriptures failed to meet; for if a text expounded a thesis concerning a point of controversy, it was almost certain that the anti-thesis would exist in another scripture. Thus, the majority of scriptures were themselves more the objects of verification than sources of it. Be that as it may, one thing is clear: that the privileged status of the Buddha as an enlightened and omniscient being did not guarantee a privileged status to his word as regarded questions of truth; and if the veracity of buddhavacanam was not post-hoc certain, then it necessitated a method for its verification. Truth and Authenticity It was the need for reconciling the divergent opinions expressed in the Buddhist scriptures that led to a new genre of texts. If the Nikdyas, the Abhidharma9 and the Prajndpdramitd represent a first order or base level of scripture, sutras such as the Samdhinirmocana, which attempt to arbitrate inconsistencies between first-order scriptures, can be termed second-order or level two meta-scriptures. By the time such questions had reached the great Tibetan master Tsong kha pa for example, the issues were at least third-order (and sometimes fourth). Tsong kha pa not only tackled the problem of reconciling two first order scriptures, he also took as his subject matter second order scriptures such as the Samdhinirmocana, trying to reconcile its claims (which he or course considered to be buddhavacanam) with those made in other sutras and sdstras. It is important to note here that Tsong kha pa conducted his analysis not so as to be considered a third-level meta-physician (in the 9 literal sense of the term), but because he saw a real need to come to grips with the problems of meaning and truth that confronted him. His analysis was not a mere intellectual game, but an earnest attempt to answer questions he felt to be soteriologically important; and to do so in a thoroughly non-dogmatic fashion. He states at the beginning of his Legs hshad snying po: It is impossible to elucidate (the status of a scripture) simply (by relying upon) a text which says "this is of direct meaning (nges don, skt. nitartha)" because (were this the case), all the commentaries composed by Mahayanists would have been pointless. Moreover, there are many disagreements between the very texts which say that they settle (the question of what is) direct and what indirect meaning. One is unable to settle the issue by simply (quoting) a scripture which says "this (text) is of such and such (a meaning)" because when it cannot be done (in this way as regards) general questions (i.e., first order questions), (why should it be so as regards) the specific issue of direct/indirect (meaning) (i.e., second order questions)? 10 He concludes that In the end, it is necessary to distinguish (such texts) by nonmistaken reasoning itself.'' and not by relying on dogma. T o sum up, then, second-order scriptures attempt to reconcile inconsistencies between first-order ones. Third-order texts deal with the inconsistencies of second-order texts, and so on. The thing to remember is that in this hermeneutical circus, the tricks become successively more and more daring as we proceed from level to level. Before we can discuss the actual modus operandi of the reconciliation of inconsistencies, one major question needs to be answered: Why the need for reconciliation, arbitration or interpretation at all? After all, if two religious texts diverge, the simplest solution is to challenge the authenticity of one of them and to claim that the historically later one is apocryphal. This attitude has existed throughout Buddhist history, but it has for the most part been one-sided. The Sautrantikas criticized the innnovative Abhidharmists. According to traditional hagiography Vasubandhu initially criticized the "heretical Mahayana" followed by his brother Asahga; and indeed, even today, we see some Theravadins 10 making the same kinds of criticisms of the Mahayana scriptures. The critique, however, is luckily one-sided, for the Abhidharmists (as far as we know) did not call into question the authenticity of the Nikdyas, nor does the Mahayana deny that the Pali canon is buddhavacanam. Indeed, it is one of the Bodhisattva's root vows to refrain from "disparaging the srdvakaydna" (nyan-smod)*2. It is quite fortunate that the debate did not for the most part center upon questions of authenticity, for excessive preoccupation with such issues could only have led to dogmatism, and to the stagnation of the tradition. T h e emergence of new scriptures and the reinterpretation of old ones is a sign of the vitality of a tradition. Thus, with the Mahayana sutras, the Tantric scriptures, and even the Tibetan dgongs gter, we find a steady influx of creativity into the tradition. T o dismiss them as apocryphal is simply to skirt the real issue, that of their meaning. Instead, it seems that once a siltra (or a tantra for that matter) had been around for a while, it became accepted as buddhavacanam, and once this occurred, it was its contents, its meaning and its veracity (and not its authorship) that became the object of debate. From that point on it was only its status as either of direct meaning (nitdrtha, Tib. nges don) or of indirect meaning {neydrtha, Tib. drang don) that came into question, and not its authenticity. Truth in the Buddhist Scriptures Buddhists have traditionally held that the word of their founder expresses the truth (satya, tib. bden pa), and now we must inquire as to the meaning of this apparently very dogmatic statement. In his Chos 'byung, the great Bu ston Rin chen sgrub quotes a sutra passage describing the Buddha's doctrine as being "of good meaning" (si>artha, Tib. don bzang-po), and he comments: '"of good meaning' refers to the perfection of the subject matter which is incontrovertible." 13 Moreover, the tenth of the sixty good qualities of the Buddha's word (sastydkara upeta vak, Tib. yan lag drug cu dang Idan pai gsung) is that it is "free from fault" 14 ; the twenty-ninth that "it is correct because it does not contradict pramdna"15 and the fifty-first that it is "perfect since it brings about completion of all the aims of beings." 16 Now given this characterization of the Buddha's word the obvious question is, can the word of the Buddha (or of great saints such 11 as Nagarjuna and Asanga) be anything but true? T h e answer comes variously. The Lahkavatdra, in its usual radical style, has this to say. T h e Mahayana is neither my vehicle nor (my) speech, nor (my) words; it is neither the truth, nor liberation nor the realm without appearances. 17 And again: Nirvana is where the idea of truth is not adhered to because it is confusing. 18 And yet, despite the fact that the Lahkavatdra de-emphasizes the importance of the notion of truth, the tradition has placed a great deal of emphasis on just such a notion. Let us turn for a moment to Tibet, and in particular to a series of debates that occurred between the eighth Kar ma pa, Mi bskyod rdo rje, and the dGe lugs scholar Se rva rje mtshun Chos kyi rgyal mtshan. In the latter's 'Gag Ian kLu sgrub dgongs rgyan, he ascribes the following position to the Kar ma pa: . . . when one is commenting on the meaning of a sutra which teaches the madhyamika view, if one interprets it as cittamdtra, it will be the ruin of the teachings (bstan pa chud gzan pa).>9 T h e work being referred to here is not a sutra but a sdstra of Vasubandhu's, and the view being expressed by the Kar ma pa is a very common-sense one. If Vasubandhu's commentary interprets the Prajhdpdramitd sutras (which both Se rva rje mtshun pa and the Kar ma pa accept as Madhyamika works) as if they were Cittamatrin, then Vasubandhu is in error, and his text cannot be said to expound the truth. But in reply, Se rva rje mtshun pa has this to say: T h e Acarya Santipa explained the intended meaning of the Prajndpdramitd Sutras to be the Cittamatra. T h e Catursatakabhdsya also says that the Sthavira Dharmapala explained the intended meaning of the Mulamddhyamihakdrikds as Cittamatrin. Now because these (sages) interpret sutras which expound {'chad) the Madhyamika view . . . as Cittamatra, were this to ruin the teachings (as the Kar ma pa claims), then (one would be reduced to saying that) similar to those two sages, the Lord (himself), in 12 his own scriptures (ruined the teachings); for (did not the Buddha himself) extensively teach the Cittamatra views as the third wheel for the purpose of leading the disciples who have tendencies (rigs) toward the Cittamatra? 20 Se rva rje mtshun-pa's point is this: to mis-interpret (whether deliberately or not) is not necessarily to ruin. A hermeneutical fallacy does not necessarily lead to a scripture's being considered false or "ruined." At this point, we might once again ask our question: what does the tradition mean when it says that the Buddha's word is true, and does asserting that it is true (in the sense with which the tradition uses the word true) preclude all possibility of its being fallacious? This latter position, that it is logically impossible for the Buddha's word to be false, is, to put it mildly, rather dogmatic. It is, as I hope I have made clear, not at all what is meant by the above claims that the word of the Buddha is true. Instead, the word "true" in the above contexts has a definite pragmatic tinge to it. When Bu-ston characterizes buddhavacanam as svartha, as being "of good meaning," when the sutras call it "perfect," "correct" and 'free from fault," or when Se rva rje mtshun pa claims that the Buddha's doctrine is valid or true despite inconsistency, they are not claiming that all of the scriptures are unconditionally true, but that they are pragmatically true. They are pragmatically true because they are all conducive to the spiritual development of those who hear them. Kajiyama hits the nail on the head when he says that "the lower doctrines were not simply rejected but admitted as steps leading to an understanding of the higher ones." 21 T h e Buddha's word is well-spoken (subhdsita), says the Vydkhydyukti, for ten reasons, the fifth one being that it is spoken "in accordance with the intellectual faculty of various human beings." We can now see quite clearly that this is what is being pointed to when the word "true" is predicated of the Buddha's word. "Truth" here refers to soteriological validity and not to the absence of logical inconsistency. With this more pragmatic sense of "truth," we can see why the tradition makes the claim, as did my Theravadin colleague, that all of the words of the Buddha are true. Scriptural Inconsistency and Its Solution I began this paper with several claims as to the non-dogmatic 13 nature of Buddhist doctrine, and yet I have thus far made two apparently very dogmatic statements: first, that not only is questioning the authenticity of scripture not important according to the Mahayana, it is in fact discouraged by certain vows;— and secondly, that the Buddha's word is in its entirety true (in the pragmatic sense described above). T o preclude debate about authenticity shifts the focus of attention from authorship (pudgala) to doctrine (dharma). To make the unqualified assertion that all of the scriptures are pragmatically true accomplishes two things. It first of all reaffirms the presupposition of the Buddha's status as an enlightened being who "never speaks without a special purpose"; 2 3 and more importantly, it engenders within the adept a sense of respect for the teachings, which he now considers relevant to his spiritual progress. It implicitly shifts the focus of attention from considering the doctrine as mere words {x>yanjana) to considering it as relevant, or full of meaning {artha).14 If the Buddhist scriptures are authentically the word of the Bhagavan, and if they are pragmatically true, then two possible means for resolving the contradictions that arise in scripture have been precluded. We can neither take the dogmatic route of dismissing scriptures as spurious; nor can we deny the perfection of the Buddha by dismissing some of his scriptures as pragmatically false, as lacking soteriological value. And now, in a state of utter despondency, we may echo the words of the Bodhisattva Don dam Yang dag 'Phags in the Samdhinirmocana, as paraphrased by Tsong kha pa in the Legs bshad snying po\ We see that in some sutras (The Lord) says that all dharmas lack svabhava, etc. In others, the svalaksana of the aggregates, etc., are said to exist. When we compare these two statements a contradiction arises, and since there should be no contradictions, I ask (7 he Lord): with what intention did you say that svabhavas do not exist? 25 T h e r e is indeed a third alternative for resolving such inconsistencies, and it comes in the form of the doctrines oi' neyartha and wtdrtha. It is neither the authenticity nor the pragmatic truth of the Buddhist scriptures which the tradition questions, but only their intended meaning (samdhi, Tib. dgongs). In short, something had to give, and if it was neither authenticity nor soteriological worth, then it had to be meaning or intention. 14 All of the scriptures had two properties in common: they were all authentically the word of the Buddha, and they were all pragmatically true. They differed in that they were not all considered to be unconditionally true, which is to say that when subjected to analysis, some were found to be faulty, though at all times soteriologically valid. Those which passed the test of critical evaluation, which were considered unconditionally true, were labeled as of definitive meaning (nitdrtha, Tib. nges don), which is to say that they were considered to be the ultimate intention (mbhar thug pa'i dgongs pa) of the Buddha. In a word, the focus changed from considering the word of the Buddha as true, to considering truth to be the Buddha's word (or at least his intention). Now the way in which this was accomplished, the method for setting up the concepts of neydrtha and nitdrtha, varied from school to school. In his discussion of these concepts in the Hinayana scriptures, Jayatilleke has this to say: "When he (the Buddha) is pointing out the misleading implications of speech . . . his meaning is direct." 26 Though this may be one interpretation of what it means for a text to be nitdrtha, it is certainly not one that would be accepted by a follower of Mahayana. Within the latter system, we have an overabundance of data regarding the doctrines of direct and indirect meaning. The issue is raised in the Lahkdvatdra, the Samdhinirmocana, and Candrakirti's Prasdnnapadd, and it becomes especially important in Tibetan exegetical literature, especially within the gZhan stong commentaries of Shakya mchog ldan and Dol bu pa, in the works of Bu ston Rinpoche, and of course in Tsong kha pa's Drang nges legs bshad snying po, which in turn has its own corpus of commentarial literature. To this latter interpretation we now turn. Scriptures of Indirect Meaning Implicitly in the Legs bshad snying po, and quite explicitly in some of his other works (such as the Legs bshad gser 'phreng), Tsong kha pa states that a text must meet three criteria to be considered of indirect meaning (drang don). These are: (1) That it have a basis of intention (dgongs gzhi) (2) That it have the property of necessity (dgos pa) (3) That it contradict reality (dngos la gnod byed) if taken literally. 15 If a treatise is to be considered of interpretive meaning, if it cannot be taken literally, then there must be some correct interpretation of the text. This is referred to as the "basis of intention." It is the actual or ultimate meaning of a text or passage. T h e r e must also be a necessity {dgospa) in its having been taught with such a concealed intention or in such a hidden fashion. This is the second criterion which a text of indirect meaning must meet. Finally, says Tsong kha pa, there must be some logical inconsistency which results from taking the passage as it stands without attempting to identify the actual meaning. Were there no contradiction (gnod—literally "harm") in taking the apparent meaning as the actual intention of the text, then the text would not be of interpretive meaning but of definitive meaning. Some examples should clarify these criteria. Again we turn to the Se rva rje mtshun-pa / Mi bskyod rdo rje debates. There we find the former scholar making the assertion that the last three works of Maitreya (The Mahdydnasutrdlamkdra,21 the Mddhydntawbhahga,2* and the Dharrnadharmatdvibhahga29) are the Cittamatra treatises (and not Madhyamika ones) because they put forth the doctrine of three final vehicles (triydnavdda, Tib. mthar thugthegpa gsum),i() interpreting sutras which teach the ekaydna as being of interpretive meaning. We find in Sutrdlamkdra XI, 53, for example, the seven "bases of intention" (dgongs gzhi) for the doctrine of the ekaydna. Since the Sutrdlamkdra expounds the doctrine of three final vehicles, it finds objectionable the doctrine of the ekaydna, and sets out to interpret it as a doctrine which cannot be taken literally (as neydrtha) by positing these seven bases of intention, which it claims to be the doctrines actually intended by the Buddha when he taught the provisional doctrine of one final vehicle. Suffice it here to cite just the second of these dgongs gzhi, nairatmya tulyavat.ix All of the vehicles are "equivalent (as regards the fact that they all teach) selflessness," and it is because of this similarity in the vehicles and not because there is ultimately one final vehicle, that the Buddha taught the ekaydna. The commentary explains: . . . that there is one final vehicle (taught) due to an equivalence as regards selflessness means that there is a similarity in the vehicles of the srdvakas, etc. as regards the non-existence of a self.316 This, then, is an example of the dgongs gzhi or the "basis of intention." It is the actual or ultimate intention of a text or passage, the basis which underlies whatever provisional doctrine is expressed by taking the text literally, the basis which constitutes the correct interpretation of the text. T h e claim being made by the Sutralamkdra is that when the Buddha taught the doctrine of the ekaydna, his actual intent (his dgongs gzhi) was to point out similarities in the tenets (such as selflessness) of the different vehicles. He did not therefore intend that the doctrine of the ekaydna be taken literally—this according to the Sutralamkdra. Again, in response to the claim that the tathdgatagarbha is a self, the Lahkdvatdra says regarding dgongs gzhi: T h e Lord spoke: my doctrine of the tathdgatagarbha, Mahamati, is not like the self-doctrine of the heretics. For the Tathagatas, Mahamati, teach the doctrine of the tathdgatagarbha having designated it to mean sunyatd.33 T h e dgongs gzhi or "basis of intention" of the doctrine of the tathdgatagarbha is, according to the Lanka, nothing but sunydta. It thus asserts that statements, such as those in the Ratnagotravibhahga, which claim the tathdgatagarbha to be a self (dtman) that is permanent (nitya), etc., could not be taken literally. We have thus seen two examples of the way in which dgongs gzhi forms an integral part of the process of classifying a work as neydrtha. Dgos pa, or "necessity," must also be present. Why was it necessary for the Buddha to teach the doctrine of the ekaydna if it cannot be taken as unconditionally true? T h e Sutralamkdra replies (XI, 54): So as to convert some and so as to hold onto others, the fully enlightened ones have taught the ekaydna to those of indefinite (potential). 34 T h e commentary goes on to explain that although there are three final vehicles, there are some beings (of indefinite potential—aniyata) who could take either Mahayana or Sravaka paths, and that the existence of these beings necessitated (dgos pa) the teaching of the ekaydna. Not to have taught it would have meant that these beings might have failed to realize their full potential. 17 Again, in regard to the tathagatagarbha there is also a dgos pa. T h e Lahkdvatdra states: T h e Tathagatas, The Arhants, the Fully Enlightened Ones teach the state of non-discrimination, the state without appearances, by means of the doctrine suggesting a tathagatagarbha so as to turn away the fear of egolessness which worldlings have. 35 Thus, according to the Lanka, it is "necessary" (dgos) to expound such a doctrine as the tathdgatabargha as an indirect teaching so as skillfully to lead those beings who fear nairdtmya to an understanding of emptiness. In other words, it is a question of updyakausalya. T h e third criterion, that there must be some fallacy (gnod) in taking these tenets as they stand, is the crucial point, for if no fallacy could be found, then the first two points would have been made in vain. T h e first two criteria, ascribing actual intention and motivation to certain teachings, can be seen more as outcomes of the third. Which is to say that where a doctrine does not contradict reality (dngos la gnod byed) there is no need to determine a basis of intention (dgongs gzhi) or a necessity (dgos pa). This then is the essence of a text of indirect meaning, that it contradict reality; and to state the contrapositive, if a text is to be of direct meaning, it cannot contradict reality. Scriptures of Direct Meaning What kind of doctrine, then, what text, does not contradict reality? Different schools of Buddhist philosophy have answered this question in different ways. Indeed, it is this fact which makes them different. According to the Madhyamika, there is only one doctrine that does not contradict reality, and that is, of course, emptiness. Thus, scriptures which teach emptiness are identified as of definitive meaning (nitdrtha, tib.nges-don) by the Madhyamika. In discussing this point, both Bu ston and Tsong kha pa cite this famous passage from the Aksyamati nirdesa: What are the sutras of definitive meaning and what the sutras of interpretive meaning? T h e sutras which teach the conventional are said to be of interpretive meaning, and those which teach the ultimate are said to be of definitive meaning. Those sutras which 18 teach various words and letters are said to be interpretive sutras. Those sutras which teach the profound, the difficult to see, the difficult to realize, those are said to be of definitive meaning. T h e sutras which teach concepts such as self, beings, life, nourishment, mankind, personality (etc.) . . . these sutras are said to be of interpretive meaning. Those sutras which teach that things are empty, without characteristic, wishless, non-compounded, unarisen, unproduced; which teach that there are no beings, no life, no personality, no owners; (in short) those sutras which teach the door to emancipation should be known as of definitive meaning. And that is why it is said 'rely on scriptures of definitive meaning and not on scriptures of interpretive meaning.' 36 This idea of defining scriptures of definitive meaning in terms of whether or not they teach emptiness seems to be a characteristic of Madhyamika thought. Still, we do find implicit statements to that effect in non-Madhyamikan works. We find in the Lahkdvatdra for example the following lines: And again, O Mahamati, the teachings of the self-nature of entities and of general characteristics are, O Mahamati, the teachings of the manifest Buddha and not the teachings of the dharmatd Buddha. 3 7 T h e Lanka goes on to identify such teachings as meant for the childish ones (bdla), thereby implicitly giving a more definitive character to the doctrine which expounds no-self-nature, i.e., emptiness. However, such passages are rare and, in contradistinction to the Aksayamatinirde'sa, the Lanka can in general be said to repudiate the notion that sutras which teach emptiness demonstrate the unqualified truth. For the Lanka, the ultimate truth is ineffable and beyond depiction by words. T h e linguistic categories of direct and indirect meaning are inapplicable to the ultimate, which is only the object of "the wisdom that a noble one has of the truth" (tattvaryajnana).™ This raises a question which is hotly debated in Madhyamikan circles as well. Given the general Buddhist belief that language is incapable of depicting reality, how can any doctrine that is expressed verbally (as sunyatd is) help but contradict reality? And if it does contradict reality (which you will recall is the principal criterion characterizing a sutra of interpretive meaning) then how can it be of definitive meaning (nitartha)} We seem to be faced with a paradox: for a scripture to be considered nitartha is must linguistically depict empti19 ness, and yet in the act of linguistically depicting, it is reduced to the level of neydrtha. T h e question is a complex one, involving, among other things, issues in the philosophy of mysticism.39 It is, however, beyond the scope of this paper. Suffice it here to say that for Tsong kha pa, the ineffability of sunyatd does not imply that it is incapable of being depicted linguistically; and it is exactly the correct enunciation of the doctrine of emptiness which characterizes a scripture of direct meaning. This is, according to Tsong kha pa, the ultimate intent of the Buddha; it is the unqualified truth. Thus, any scripture which fails to teach emptiness must of necessity be interpreted. How is one to know which conception of sunayatd is the right one? Tsong kha pa answers: "through non-mistaken reasoning itself." Thus, in the end, it is the critical spirit which triumphs. If along the way spatio-temporal concerns such as authenticity are disregarded; and if overtly religious presupposition (such as the infallibility of the Buddha) prohibit the repudiation of the pragmatic value of the doctrine, it is only to pave the way for the truly important questions, those of the truth, and hence of the ultimate intent of the Buddha's scriptures. In the end, it is not so much that the words of the Buddha are true, as it is that the enunciation of ultimate truth becomes the sole criterion of the Buddha's intention. NOTES 1. This paper is a revised and enlarged version of a paper that was read at the conference of the International Association of Buddhist Studies held jointly with the International Association for the History of Religions in Winnipeg Canada on August 19, 1980. T h e author wishes to express his thanks to a number of very learned kalydnamitras from whom he has been fortunate enough to receive instruction. First and foremost is Geshe Lhundub Sopa with whom the author has had the good luck to read major portions of the Drang nges legs bshad snying po of Tsong kha pa and the dGag tan hLu sgrub gdongs rgyan of Se rva rje mtshun pa. Thanks must also go to Kensur Wangdu Rinpoche, the former abbot of the Lower Tantric College, Samdong Rinpoche, Director of the Tibetan Institute of Higher Studies in Sarnath, Geshe Wangchen. tutor in Tibetan Studies for the University of Wisconsin Year in India Program, and Geshe I.obsand Tseringof Se rva Byas Monastic College in India for some very lucid explanations and discussions on the meaning of pdyad pa gsum gyis dag pat lung (see below). Finally my thanks go to Mr. Roger Jackson and Gelong Thubten Thardo for reading the manuscript of the paper and making valuable suggestions for improvement. Of course, the views, interpretations, and errors expressed herein are ultimately the author's own. 2. I am here particularly thinking of Ksho Mikoganii's article "The Problem of 20 Verbal Testimony in Yogacara Buddhism," Bukkyotaku knikyu No. 32 and 33, 1977, in which he seems to ascribe to the Yogdcdrabhumi a quite rigid dogmatism. 3. The translation requires some justification. I have translated the word sdmdidfl (literally "due to its similarity" or "due to generality") by the word "inductive" because this is quite clearly what the commentary takes it to be. It (the reliable script ure) can be inferred to be accurate as concerns very subtle points because of its similarity with other scriptural points, less subtle, which can be determined to be accurate by either deductive reasoning (dngos stobs rjes dpag) or by direct perception (mngon sum), h is because those other, more evident, points are determined to be correct, that one can infer that the extremely subtle points (shin hi Ikog gyur) are accurate. This is clearly a case of inductive reasoning. 4. The Pramdriavartikam of Dharmakirti, R. Gnoli, editor, Serie Orientalie Roma, Rome, 1960, p. 109. This is one of the most controversial verses in Buddhist pramdna literature, especially in Tibet. In the near future, I plan to devote an entire paper to the elucidation of the different traditions of interpretation of this verse, both Indian and Tibetan. The verse originally appears in Pramdnasamuccaya I!, 5. 5. rGyal tshab Dar ma rin chen quotes this verse in Thar lam gsal byed, Pleasure of Elegant Sayings Press, Sarnath, 1974, p. 175. 6. T h e commentary on this point inns from p. 175 to p. 180 of the Sarnath edition (see previous note). 7. Whether induction in such a case is warranted is, I think, an open question. In any case, the most that one can hope to achieve from an inductive argument is that the conclusion is likely (that the very subtle points are most likely accurate) and not that it is certain. From my conversations with several scholars of the tradition, this seems to be not altogether appealing. 8. Again, there is an abundance of variant interpretations within the dGe lugs pa tradition alone regarding this third criterion. A more precise way of putting it, as rGyal tshab rje does, is to say that the scripture must be internally consistent. The former parts cannot contradict the latter (snga spyi gal ba med pa), nor can the explicit meaning contradict what is implied (dngvs \hugs gal ba med pa). 9. R. T h u r m a n , in an excellent article entitled "Buddhist Hermeneutics" (Journal of the American Academy of Religion, XLV1/I p. 25), suggests that the Abhidharma itself "contains the earliest forms of the hermeneutical concepts," and this can certainly be agreed to provided that we make a distinction between a synthetic hermeneutic which attempts to synthesize analogous doctrines into a logical whole, and a didactic hermeneutic which attempts to reconcile contradictory doctrines by interpretation. T h e former is first order, the latter at least second. It seems to me that the Abhidharma is of the synthetic (and therefore first older) variety. 10. Geshe I. Rabtaii, Drang ages main 'Iryed legs bsliad snying pu dka' gnad mams mchan bur bkod pa gzur gnas bto gsal dkai ston, (annotations on the Letgs bshad syning po of I song kha pa with the root text), I.hun grub chos grags, Delhi, p. 5. 11. Ibid. p. 5. 12. See, for example, the list of Bodhisattva vows in Pha bong kha Rin po che's Thun drug gi riutl 'hyor, found in bl.a ma'i rnal ~b\mS/ie.s rig par kh/mg, Dharmasala, p. 28. 13. Bu ston rin chen grub, Collected Works, edited by Lokesh Chandra, New Dehli, 1956-1971. In the volume of his Chos 'hung, p. 677. 14. Ibid. p. 651. 15. Ibid. pp. 6 5 2 - 3 . \6.Ibid. p. 654. 17. Lankdvatdra sdtra, F.L. Vaidya, editor, Buddhist Sanskrit Texts No. 3; Mithila Institute, Darbhanga, 1963, p. 56. 18. Ibid. p. 75. J 9. Se rva rje mtshun Chos kyi rGyal mtshan, 'dag Ian kLu sgrub dgongs rgyan, Tibet House, New Delhi, 1969, p. 9. 20.1bid. p. 18. 21. In M. Kiyota's Maliayana Buddhist Meditation U. of Hawaii Press, Honolulu, 1978, p. 117. 22. See note 12. 23. Bu ston, op. fit. p. 653. 24. As per the famous saying: arthapratisaraneria bhavitaiyamna lyahjanupratisaraneria. dhurmapratisaranenu bhavitavyam na pudgalapratisaranena (etc.). In the Mahdtyutpatti the four pradsumvidah are given <Cf. XIII, 196). 25. Tsong kha pa, op. (it., p. 11. 26. K. N. Jayatilleke, Early Buddhist Theory of Knowledge, Motilal Banarsidass, New Dehli. 1963, p. 363. 27. Ho L'i, et al., A Complete Catalogue of the Tibetan Buddhist Canons, Tokyo Imperial University, Japan, 1934, p. 609, No. 4020. 28. Ibid. p. 609, No. 4021. 29. Ibid. p. 609, No. 4022. 30. I.e., that there are three separate and distinct results that are the fruits of the Mahayana and Hinayana practices. It is a specific feature of certain subschools of the Yogacara to claim that there are three possible ultimate fruits of the Buddhist path—the enlightenment of a 'srdvaka, the enlightenment of a pratyekalmddfia, and that of a fully-enlightened Buddha. This claim is a feature of these Yogacara schools which distinguishes them from the Madhyamika. 31. Mahdyanasutrdlarhkdra, P.L. Vaidya editor, Buddhist Sanskrit Texts No. 13, Mithila Institute, Darbhanga, 1963, p. 67. 32. Ibid. p. 67. T h e Sanskrit reads: nairdtmya tutyatvdd ekaydnatd srdvakddindmdtmabhavatd samanyddyatd ydtiamiti krh'd . . . 33. Lahkdvatdra, p. 32. Bhagavdndha—na hi mahdmate tirthikdtmavddatulyo mama tathdgalagarbhopade'mh. kirn tu mahdmate tathdgatah sunyatd . . . piuiarthdnam tathdgatagarbhopadesamkrtvd . . . de'sayanti. 34. Sutrdlamkdra, p. 69. 35. Lahkdvatdra, p. 33. tathdgatd arhantah samyaksambuddhd bdldndm nairdtmyasamtrdaapadavivdrjanartham nirvikalpamrdbhasagocaram Uithdgatagarbliamukhopade'sena de'sayanti. 36. As it appears in Geshe T. Rabten's l.egs-bshad snying-po commentary (see note 10) p. 184. 37. Lahkdvatdra, p. 3. ya punareva mahdmate bhdvasvabhdva sdmdnya taksana desand esa mahdmate nairmanikabuddha de'sann, na dhannatd buddha dAanu. 38. Ibid. p. 33. 39. The whole question of ineflability as discussed by W. Stace in his Philosophy of Mysticism is quite relevant here. It is my belief that Professor Stace's arguments 22 against the cogency of the claim of ineffability are simply inapplicable to the Buddhist conception of what it means for something to be ineffable. 23